A vehicle lender earns a 15% yield, pays 9% on borrowings of five sixths of its assets, spends 3% on operations and loses 1.5% to credit costs. Compute ROA and ROE, then find what each single lever must do to reach a 16% ROE.

ROA is 2.25% and ROE 13.5%; reaching 16% needs about 0.56 points more of pre-tax ROA, Rs 111 crore a year, from any one lever. That is yield up to 15.56%, cost of funds down to 8.33%, opex down to 2.44%, credit cost down to 0.94%, or leverage up to 8.2x. In points they look alike; as a share of each line they do not. Credit cost would have to fall 37%. The board should rely on funding cost and pricing, and treat credit cost as the lever that moves on its own.

Think of a shop that borrows most of its stock money from a wholesaler. Its profit on sales is thin, but because very little of its own money is in the shop, the return on that money can be high. A lender's ROE is a thin return on assets multiplied by leverage, so every line is measured in fractions of a per cent of assets and every fraction is worth six times as much to equity. Interest is the line people get wrong: 9% applies only to the five sixths that is borrowed, so it costs 7.5% of assets. Yield 15% less 7.5% leaves a net interest marginInterest earned on assets less interest paid on funding, as a share of assets. of 7.5%; opex of 3% and credit cost of 1.5% leave 3.0% before tax.

Work backwards from the target. A 16% ROE at six times leverage needs an ROA of 2.667%, which is 3.556% before tax. The gap is 0.56 points of pre-tax ROA, about Rs 111 crore a year on Rs 20,000 crore of assets, and each lever must supply all of it alone. Yield must rise to 15.56%. Cost of funds must fall by 0.67 points, to 8.33%, because it applies to only five sixths of assets. Opex must fall to 2.44% and credit cost to 0.94%. Leverage is different: more borrowing adds interest cost as well as scale, so ROE rises by only 1.125 points for each extra turn, and reaching 16% takes 8.2x.

Rank them by how much of the move is in management's hands. Funding cost and pricing are the credible levers; credit cost is the one that moves on its own, in both directions. A 0.67 point cut in borrowing cost is what a rating upgrade or a shift from bank lines to bonds typically aims at, and a 0.56 point rise in yield is a pricing or mix decision, though higher-yield segments usually bring higher credit cost with them. Cutting a fifth of operating cost from a business that collects instalments in cash across small towns is slow. Leverage of 8.2x may sit above what capital rules or lenders allow, so check the current regulatory minimum before offering it. And credit cost works the other way just as easily: a bad year that lifts it by the same 0.56 points, to 2.06%, takes ROE down to 11.0%. That asymmetry is why rating analysts watch a lender's credit cost line before any other.